accurate timer ..would an rtc with a good oscillator do?
rwgast_logicdesign
Posts: 1,464
ok so i have wanted to be able to do some accurate things as far a pulses and frequency counting for a really long time. now lets say you wanted to determine what freq your prop is running at by counting pulses. do do this you would probably want to start the pulsing for a full second and use the the counters or maybe even an external 4060 or the like. the problem if you need a something with good accuracy to start and stop at one second.
now i know without some specialized 20 dollar plus oscillaters your not gunna get dead on but pretty close is good enough as in 1000hz plus minus.
what im wondering is how accurate would a ds1302/1307 to be the stop watch basically? i know alot of error comes from the xtal so ive eliminated that buy using a ds32khz tempature compensated oscillator, its data sheet says its plus or minus 4 seconds a year.. although i dont see any ppm number. would this set up be highly accurate for starting stopin processes in short second and milla second time spans? if not exaxtly what would be better thats not a rediculous price maybe an lm322 or 555 with low tolerance passives or a 74hc4060.
the reason i like the rtc is becuase it usefull in lots of other ways adding an rtc to most systems is a need anyways. I just see alot of timer counter type chips out there that a think an rtc could do.
now i know without some specialized 20 dollar plus oscillaters your not gunna get dead on but pretty close is good enough as in 1000hz plus minus.
what im wondering is how accurate would a ds1302/1307 to be the stop watch basically? i know alot of error comes from the xtal so ive eliminated that buy using a ds32khz tempature compensated oscillator, its data sheet says its plus or minus 4 seconds a year.. although i dont see any ppm number. would this set up be highly accurate for starting stopin processes in short second and milla second time spans? if not exaxtly what would be better thats not a rediculous price maybe an lm322 or 555 with low tolerance passives or a 74hc4060.
the reason i like the rtc is becuase it usefull in lots of other ways adding an rtc to most systems is a need anyways. I just see alot of timer counter type chips out there that a think an rtc could do.
Comments
So you need a DS32KHz or a DS3232/DS3232M
If you just want to check on the bench, the 1pps from a GPS module is simple and precise.
Or, Something like ASVTX-09-20.000MHZ-T, will give you 0.5ppm, VCTCXO
Be a little cautious with 'corrected' RTC, as some fudge things a little and use a swallow counter, so you need to ensure you are whole-trim-cycle , otherwise you will get significant jitter.
For a RTC which just needs time of day, this is tolerable.
The PCF2129A is also quite well spec'd, says Typical accuracy: ±3 ppm from −15 °C to +60 °C and also this
Temperature compensated crystal oscillator
The frequency of tuning fork quartz crystal oscillators is temperature-dependent. In the
PCF2129A the frequency drift caused by temperature variation is corrected by adjusting
the load capacitance of the crystal oscillator.
The load capacitance is changed by switching between two load capacitance values using
a modulation signal with a programmable duty cycle. Every chip is calibrated in order to
produce, at the measured temperature, the correct duty cycle which compensates for the
frequency drift.
The frequency accuracy can be evaluated by measuring the frequency of the square
wave signal available at the output pin CLKOUT. However, the selection of
fCLKOUT = 32.768 kHz (default value) leads to inaccurate measurements. The most
accurate frequency measurement occurs when fCLKOUT = 1 Hz is selected (see Table 10).
By correction modulating the XTAL-CAP, the jitter is reduced, but even there you need to be 1s or multiples there of, to get full correction cycles and best precision.
I just purchased a PCF2127. It's just .60 more and has some extra benefits.
Page 3 of this document.
Well spotted. Adds RAM / Power Fail / Reset - same Tempco.